The hardest part of an AI project isn't the AI.
In my experience, the more important question is whether the problem has been understood well enough that data or computational methods can genuinely contribute.
Much of my career—as a researcher, educator, and builder of interdisciplinary programs—has focused on helping people with different forms of expertise work together to answer exactly that question.
Whether the challenge involves scientific research, healthcare, education, public policy, or commercial innovation, successful projects rarely depend on technology alone. They depend on creating a shared understanding among the people who know the problem, the people who understand the available data, and the people who develop computational solutions.
That is where I can help.
Approach
Clear thinking before complex machinery.
For more than three decades, my research has examined how people learn, represent knowledge, communicate, and make decisions, alongside the development of computational models intended to explain those processes.
That perspective shapes how I approach organizational problems involving data and AI. The work often begins not with a model, but with a conversation: What decision needs to be improved? What uncertainty needs to be reduced? What would count as success? What expertise is missing from the room?
The goal is not to force every challenge into a computational form. It is to determine where technical methods can make a meaningful contribution, and to create the shared understanding required for those methods to be used well.
How I typically contribute
Helping people with different forms of expertise think together.
The most useful role is often not that of a conventional technical vendor. It is to help shape the problem, translate across perspectives, and create a sound basis for action.
01
Clarifying difficult problems
Many organizations begin with a solution in mind. I help identify the underlying question that actually needs to be answered, the decision that needs to be improved, and the evidence that would make progress visible.
02
Translating across expertise
Domain experts, stakeholders, data scientists, engineers, and leaders often bring different assumptions and vocabularies to the same challenge. I help them develop a shared understanding of both the problem and the possible solutions.
03
Identifying meaningful opportunities for AI
I help distinguish genuine opportunities from technological enthusiasm by considering the available data, practical constraints, human context, and the criteria by which success should be judged.
04
Compiling rigorous evidence
I advise on study design, quantitative methods, model evaluation, and experiments that can determine whether a proposed approach is actually improving outcomes rather than merely producing an impressive demonstration.
05
Providing an independent technical perspective
I offer an experienced external view on project plans, analytical claims, machine-learning methods, neural networks, statistical models, behavioral data, neuroimaging, and the interpretation of complex evidence.
06
Building long-term capability
I help organizations develop the people, training, collaborative structures, and institutional knowledge needed to use data science and AI thoughtfully over time.
Why this perspective?
A career at the intersection of human intelligence, artificial systems, and interdisciplinary work.
My consulting draws on more than three decades of research in cognitive science, cognitive neuroscience, and computational modeling. Throughout that time I have been interested in a common question: how do intelligent systems—human or artificial—represent knowledge, learn from experience, and make effective decisions?
As my research evolved, so too did my interest in creating environments where people from different disciplines could work together on increasingly complex problems.
Founder and Director
LUCID
Learning, Understanding, Cognition, Intelligence, and Data Science
I created and directed LUCID at the University of Wisconsin–Madison, an interdisciplinary graduate training program bringing together researchers from psychology, neuroscience, computer science, engineering, statistics, medicine, and other fields.
The program was built around the idea that useful data science depends on both methodological sophistication and deep understanding of the substantive problem. Through training, mentoring, workshops, applied projects, and cross-disciplinary collaboration, LUCID created a setting in which people with very different forms of expertise could reason together.
That lesson continues to guide my consulting work: translation and shared understanding are not peripheral to technical success. They are often its precondition.
Creator and Faculty Program Director
Data Science in Human Behavior
Interdisciplinary training for data-rich questions about people
I also created and directed the Data Science in Human Behavior master's program, designed to prepare students to combine programming, statistics, machine learning, and behavioral science in addressing problems arising in research, industry, government, and nonprofit organizations.
The program reflects a broader philosophy: the most interesting problems are rarely solved by a single discipline. Progress comes from bringing together people with different forms of expertise and creating the conditions under which they can think effectively together.
The conversations I enjoy
My work often begins with questions like these.
We're hearing a great deal about AI. Which of these technologies actually matter for us?
Our technical team and our domain experts seem to be talking past one another. How do we develop a shared understanding?
We have years of accumulated data. What questions can it realistically answer?
How should we evaluate whether an AI system is actually improving decisions?
How do we build an interdisciplinary team that can tackle this problem effectively?
About
Tim Rogers
I am Professor Emeritus at the University of Wisconsin–Madison with 30 years of experience studying human and artificial intelligence.
My research has examined semantic memory, language, conceptual knowledge, cognitive neuroscience, artificial neural networks, and computational models of human cognition. Alongside this work, I have developed interdisciplinary research programs, graduate training initiatives, and collaborations spanning the behavioral sciences, medicine, engineering, computer science, and statistics.
That work has been cited more than 20,000 times, and in 2025 I received the Jeffrey L. Elman Prize from the Cognitive Science Society for scientific contributions and community building—the same combination this consultancy is built around.
Through Semantic Consulting LLC, I bring this experience to organizations seeking thoughtful, evidence-based guidance on complex problems involving data, computation, and human expertise.
Every organization has problems that are difficult to define, difficult to measure, or difficult to solve.
Those are usually the interesting ones.
If you would like to discuss one of yours, I'd be delighted to talk.
tim@timrogers.net